TNFα-induced altered miRNA expression links to NF-κB signaling pathway in endometriosis

Endometriosis is a common gynecological inflammatory disorder characterized by immune system dysregulation, which is involved in lesion initiation and progression. Studies have demonstrated that several cytokines are associated with the evolution of endometriosis, including tumor necrosis factor-α (TNFα). TNFα is a non-glycosylated cytokine protein with potent inflammatory, cytotoxic, and angiogenic potential. In the current study, we examined the ability of TNFα to induce dysregulation of microRNAs (miRNAs) linked to NFkB-signaling pathways, thus contributing to the pathogenesis of endometriosis. Using RT-QPCR, the expression of several miRNAs were quantified in primary cells derived from eutopic endometrium of endometriosis subjects (EESC) and normal endometrial stromal cells (NESC) and also TNFα treated NESCs. The phosphorylation of the pro-inflammatory molecule NF-κB and the candidates of the survival pathways PI3K, AKT and ERK was measured by westernblot analysis. The elevated secretion of TNFα in EESCs downregulates the expression level of several miRNAs significantly (p < 0.05) in EESCs compared to NESC. Also treatment of NESCs with exogenous TNFα significantly reduced the expression of miRNAs in a dose-dependent manner to levels similar to EESCs. In addition, TNFα significantly increased the phosphorylation of the PI3K, AKT, ERK, and NF-κB signaling pathways. Notably, treatment with curcumin (CUR, diferuloylmethane), an anti-inflammatory polyphenol, significantly increased the expression of dysregulated miRNAs in EESC in a dose-dependent manner. Our findings demonstrate that TNFα is upregulated in EESCs, which subsequently dysregulates the expression of miRNAs, contributing to the pathophysiology of endometriotic cells. CUR effectively inhibits the expression of TNFα, subsequently altering miRNA levels and suppresses the phosphorylation of AKT, ERK, and NF-κB.


Background
Endometriosis is a benign, estrogen-dependent in ammatory disease characterized by the presence of endometrial tissue (speci cally glands and stroma) outside of the uterus [1] [2] [3].The exact causes of endometriosis remain unknown.The theory of retrograde menstruation, an e ux of menstrual blood and cells via the fallopian tubes to extrauterine sites, is considered an important origin of endometriosis lesions [1] [3].While 90% of reproductive-aged women experience retrograde menstruation, only 10% are diagnosed with endometriosis [4].Therefore, in addition to retrograde menstruation, other factors are likely involved in the pathogenesis of endometriosis, including hormonal imbalance, metabolic environment, epithelial-mesenchymal transition, altered immunity, and abnormal regulation of in ammation in endometrial cells (ECs) of genetically susceptible women [3].In the peritoneal cavity, resident or recruited immune cells secrete excessive levels of proin ammatory cytokines that trigger in ammatory reactions in endometrial cells and promote lesion development and disease progression [1] [5] [6] [7].

The TNFα dependent regulation of the expression of miRNAs associated with endometriosis in eutopic
ECs is not well de ned.Based on the proin ammatory nature of the disease, combined with the published data [18] [19] [22] [16] [41] [42] [49] and our comparative nanostring analysis of miRNAs (unpublished) between EESC and NESC, we aimed to analyze whether up-regulation of TNFα expression in the eutopic stromal cells of endometriotic patients induces the dysregulation of miRNAs linked to NF-kB-signaling pathways thus contributing to the pathogenesis of the disease.To evaluate this theory, the expression levels of proin ammatory and proangiogenic miRNAs were compared between the stromal cells of women with (EESC) and without endometriosis (NESC).Moreover, the effect of exogenous TNFα was tested on the expression of those selected miRNAs in NESC and whether their altered expressions are linked to the phosphorylation of NF-κB, PI3K, AKT, and/or ERK1/2 pathways.Our previous studies established that curcumin, a natural medicinal Asian herb with strong anti-in ammatory and antioxidant properties, attenuates proangiogenic and proin ammatory factors in human eutopic endometrial stromal cells through the NF-κB signaling pathway.In the current study, we further evaluated the effect of curcumin in altering the expression of proin ammatory miRNAs linked to the NF-κB signaling pathway.Taken together, we established that TNFα is upregulated in EESCs which subsequently increases the expression of proangiogenic and proin ammatory miRNAs, potentially contributing to the pathophysiology of endometriotic cells.We have determined that CUR effectively reduces the expression of TNFα and dysregulation of miRNA levels, and attenuate the phosphorylation status of PI3K, AKT, ERK, and NF-κB pathw,ays.

Human subjects and tissue acquisition
The details about the source of primary endometrial stromal cells (ESCs) used in this study were described previously [16].The current studies were approved by the institutional review boards of Emory University and Morehouse School of Medicine, Atlanta.

Endometrial stromal cell (ESC) cultures
Primary endometrial stromal cells (ESCs) from human eutopic endometrial biopsies from women with (EESC) and without evidence of endometriosis (NESC) were prepared according to [50].Cells (passages 3-5) were cultured and routinely maintained in Dulbecco's Modi ed Eagle's Medium/Ham's Nutrient Mixture F-12 (DMEM/Ham's F-12; Life Technologies, Inc.-BRL) supplemented with 12% fetal bovine serum (FBS; Thermo Fisher Scienti c, Grand Island, NY, USA), 1% non-essential amino acids, 1% sodium pyruvate, and 1% penicillin-streptomycin (Penstrep, Sigma-Aldrich, St Louis, MO, USA), within a 5% CO 2 atmosphere at 37 C in a humidi ed incubator.Cells were grown to 80% con uency in 100 mm plates (Corning, NY, USA).The culture media was replaced with low serum-containing media overnight before any experiments.After 24h, cells were treated or untreated in the DMEM/Ham's F-12 medium supplemented with 0.4% FBS, 1% non-essential amino acids, 1% sodium pyruvate, and 1% Penstrep, and incubated at 37 C in a humidi ed incubator with 5% CO 2 for 24h.Images of ESCs cultures were taken at 24 and 48 hours posttreated or untreated condition using an inverted phase contrast microscope.Unless speci ed differently, 20 random phase contrast images were acquired per well at 200× magni cation.

TNFα treatment of normal endometrial stromal cells (NESCs)
NESCs were grown up to 80% con uency in 100 mm plates as described above.Cells were serum-starved for 24 hours and then treated with TNFα (10 and 50 ng/ml, Sigma-Aldrich, USA) for 24 hours.The dose and time of treatment for TNFα are based on our unpublished work and published literature [51].Cells were harvested for the estimation of total RNA and protein.

Curcumin (CUR) treatment of normal and eutopic endometriotic stromal cells (NESCs, EESCs)
NESC and EESC cultures were grown to 80% con uency in 100 mm plates, as described above.Cells were treated with CUR (molecular weight 368.41, purity 99%, Sigma-Aldrich, USA) at a concentration of 5 and 10 µg/ml for 48h [16].CUR was dissolved in dimethylsulfoxide (DMSO) and diluted to the desired concentrations in DMEM/Ham's F-12 media with 0.4% serum-containing media followed by sterilization through 0.22µm membrane ltration.Cells were treated with the equivalent concentrations of DMSO added to the medium for the parallel vehicle control experiments.The nal concentration of DMSO was less than 0.1%.

Isolation of total RNA
Total RNA from NESC and EESC and corresponding curcumin or TNFα-treated ESCs was extracted using Qiagen miRNeasy Mini kit (Germantown, MD, USA) according to the manufacturer's instructions.The quality of the extracted RNA was veri ed via absorbance measurements at wavelengths of 230, 260, and 280 nm using a spectrophotometer (NanoDrop, 2000; Thermo Fisher Scienti c, Inc., Waltham, Massachusetts, USA).RNA 260/280 ratio of 1.9 or greater and 260/230 ratio of 1.8 or greater were used to obtain optimal results for the miR analysis.

microRNA (miR) expression analysis
The RNA samples were transcribed using the miRCURY LNA RT kit (Germantown, MD, USA) according to the manufacturer's protocol.Real-time PCR was performed using miRCURY LNA SYBR® Green PCR Kit (Germantown, MD, USA) and LNA-enhanced and Tm-normalized miRNA primers from Qiagen on CFX connected Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA).All steps were performed according to the Qiagen MicroRNA assay protocol (Germantown, MD, USA).The relative expression of the gene was calculated using 2 −ΔΔCT methods with 5S rRNA (hsa) and U6 snRNA (hsa), as the reference miRs.

Assessment of TNFα in secretion media
TNFα was measured in postculture media collected at 24 and 48 hours using Bio-Plex ProTM Human Cytokine, Chemokine, and Growth Factor Magnetic Bead-Based Assays (BioRad, Hercules, California, USA) coupled with the Luminex 200™ system (Austin, TX, USA) according to the manufacturer's protocol.Samples were tested at a 1:2 dilution using optimal concentrations of standards and antibodies according to the manufacturer's protocol.

Western blot analysis
Total protein was extracted from different treatment conditions from untreated and treated NESC and EESC and subjected to one-dimensional gel electrophoresis and western blot (WB) analysis [52].For onedimensional gel electrophoresis, equal amounts of protein (25 µg) were applied to each lane.Primary antibodies were used as described in Table 1.Membranes were incubated with the appropriate secondary antibodies for 1 hour at room temperature, and protein-antibody complexes were visualized using SuperSignal™ West Pico detection reagent (Thermo sher scienti c, Waltham, MA) on an iBright™ FL1500 Imaging System (Thermo sher scienti c, Waltham, MA).Results of representative chemiluminescence were scanned and densitometrically analyzed using a Power Macintosh Computer (G3; Apple Computer, Cupertino, CA) equipped with a Scan Jet 6100C Scanner (Hewlett-Packard, Greeley, CO).Quanti cation of the scanned images was performed using NIH Image version 1.61 software (NIH, Bethesda, MD) (34).

Statistical analysis
Data are expressed as mean ± SEM of three independent experiments.Statistical analysis was performed by one-way ANOVA using SPSS version 11.0 software (SPSS, Chicago, IL) to test the signi cance of differences in dose, duration, and interaction between dose and duration.Post-hoc corrections for multiple comparisons were done by Newman-Keuls' test or unpaired Student's t-test.Differences were considered signi cant at P ≤ 0.05.For miR expression analysis, fold change was calculated using a selected miR expression in a target sample relative to a control sample, normalized over a reference miR.The 2-∆∆Ct method was used and the ∆∆Ct was calculated using the average of the control values.
That generates multiple values close to 1 for the control and gives a standard error of the mean.

EESC secrete higher concentrations of TNFα
We compared the secretion of TNFα in the culture media of serum-starved NESC and EESC in vitro.
TNFα is known to activate the PI3K/AKT pathway, which in turn activates the NF-κB signaling pathway [66] and are essential steps for proin ammatory gene expression.So we explored whether TNFαtreatment affects phosphorylation of PI3K, AKT, ERK, and NF-κB in NESC.As shown in Figs.4A and 4B, the treatment of NESCs with TNFα at 50 ng/mL for 24h signi cantly increased the phosphorylation of PI3K, AKT, ERK, and NF-κB, whereas no signi cant effects on phosphorylation were noted at lower concentrations of TNFα except the phosphorylation of PI3K that is signi cantly higher in lower dose of TNFα.

Curcumin treatment inhibits TNFα secretion and alters the expression of miRNAs
To determine whether CUR treatment modulates the expression of miRNAs, EESCs, and NESCs were treated with different doses of CUR for 48hrs.To understand the mechanism better, TNFα secretion was analyzed post-CUR treatment.As shown in Fig. 5A, CUR treatment inhibited signi cantly the secretion (p ≤ 0.05) of TNFα in a dose-dependent manner in EESCs.In subsequent studies, the expression of selected miRNAs was analyzed under these experimental conditions.As shown in Fig. 5B, CUR treatment signi cantly promoted the expression of selected miRNAs, precisely at 5 µg/mL (miR-146a-5p) and at 10 µg/mL (miR-132-3p, miR-23a-5p) at 48 hours in EESC compared to NESC.Moreover, there is a downregulation of miRNA expression after 48 hours (miR-152-3p, miR-181a-5p, miR-199a-5p, miR-214-3p) at 5 µg/mL dose.However, there were no signi cant differences in the expression of most of the miRNAs in post-CUR-treated EESCs compared to NESCs at 48 hours (Fig. 5B).
Studies have demonstrated in endometriotic and other cells and tissues that miR-125b is involved in cell proliferation and migration [77] Similarly, miR-21 plays an essential role in the resolution of in ammation by negative feedback of in ammatory pathways [91]; [33], miR-214-3p inhibits the proliferation, migration, and invasion of EC cells [58], miR-222-3p promotes proliferation, proangiogenesis, and invasion [92]; [93], [62], and miR23a is involved in local steroidogenesis-dependent in ammation and growth of ectopic ECs [94]; [95].MicroRNA-29b is involved in a wide range of functions, including apoptosis, cell proliferation, invasion, adhesion, metabolism, and progression in endometrial cancer cells by direct regulation of PTEN [96]; [97] [98].MiR-98 expression was found to be reduced in diseased EC tissues compared to normal tissues [99].
Our results further indicate that TNFα stimulation of NESCs dysregulates miRNA expression, phenocopying EESC and implying that these cells are TNFα responsive, with effects more pronounced at higher concentrations.These ndings are consistent with previous studies indicating that a higher concentration of TNFα for a more extended exposure period promotes dysregulation of miRNAs expression, which may partly govern NF-κB-signaling molecules [64] [24] [65].Moreover, we found that exogenous TNFα signi cantly downregulated several miRNAs in NESCs except for 146a-5p which was upregulated with TNFα treatment at both doses (10 and 50 ng/mL) and miR-199a-5p, which was upregulated at the higher dosage (50 ng/mL) after 24 hours.This apparent discrepancy could be a compensatory upregulation induced by a very high concentration of exogenous TNFα for an extended period or could be a part of a negative feedback loop reducing the impact of TNFα [90].Furthermore, exogenous TNFα-dependent activation of PI3K In conclusion, the current study provides new insights into how elevated levels of TNFα secretion are associated with aberrant expression of miRNAs in ECs, which subsequently alter phosphorylation of the proin ammatory molecule NF-κB and survival pathways.Moreover, CUR treatment modulates the dysregulation of miRNAs.Further studies are needed using genetic gain or loss-of-function models of individually selected miRNAs to pinpoint the pathophysiological effects of those miRNAs in in ammation during endometriosis.Based on the dynamic nature of miRNA expression combined with diverse actions and multiple targets of NF-κB-signaling molecules, we believe that an NF-κB-miRNA feedback loop should be considered in in ammatory responses and initiation, progression, and development of endometriosis.Moreover, understanding the intersection of NF-κB signaling molecules and miRNA regulatory networks may offer opportunities for pharmacological exploitation and personalized treatment for endometriosis pain management.

Figure 1 Analysis
Figure 1

Figure 2 Analysis
Figure 2

Figure 4 The
Figure 4

Table 1
List of antibodies used for Western blot (WB) analysis.
[103]ERK1/2-signaling and NF-κB phosphorylation in NESCs suggest that TNFα may be an important cytokine contributing to the cascade of kinase signaling with dysregulation of miRNAs expression in ECs.Previous studies also established that TNFα-mediated activation of the PI3K/Akt and the NF-κB signaling pathway are essential steps for proin ammatory gene expression[66].In endometriotic cells, NK-κB signaling is activated by TNFα[12][100] [101][47]and the aberrant activation of NF-κB signaling leads to chronic in ammation, increased cell proliferation, and survival of ECs in endometriosis[102][23][24][18][19] [20],[21][22].Previous studies have also demonstrated that the phosphorylation states of NF-κB signaling molecules, including IKKα, IKKβ, NF-κB, JNK, and STAT3, are higher in EESCs, which are involved in the downstream participation of various kinases linked to cytokine-and chemokine-speci c membrane receptor complexes and adaptor proteins, that converge on NF-κB signaling pathway [16] [76][103].Thus, TNFα dependent dysregulation of miRNA expression in conjunction with altered phosphorylation of pPI3K/pAKT/pERK1/2/pNF-κB suggests a regulatory link that supports the idea of transformation of NESCs to a pathophysiological state similar to that of EESC.Further studies revealed that CUR is a potent inhibitor of TNFα secretion from EESCs[16].Moreover, our data showed that curcumin treatment could modulate TNFα mediated dysregulation of miRNAs in EESCs.The inhibitory effect of CUR is extended further to the attenuation of IKKα, IKKβ, and NF-κB[41][16] [76] [104][103].IKKα and IKKβ are part of a multiprotein complex mediating the transcription of multiple chemokine and cytokine genes through Ikβ.Thus, our results are consistent with published reports showing that CUR has strong anti-in ammatory and antiangiogenic properties[16].